EPVDemo

Contributors:

• Dan Cervone
• Alex D'Amour
• Luke Bornn
• Kirk Goldsberry

This repository contains data and code offering a demo of the NBA Expected Possession Value model presented in the paper "A Multiresolution Stochastic Process Model for Predicting NBA Possession Outcomes".

The main document that introduces and illustrates the code/data is EPV_demo.pdf. The source .tex for the tutorial file EPV_demo.pdf can be built from EPV_demo.Rnw using RStudio, or the command

Rscript -e "library(knitr); knit('./EPV_demo.Rnw')"

This operation requires the knitr R package, and takes about 1 minute on most systems. Users should be running the latest version of R with the following packages (and all of their dependencies) installed: sp, knitr, INLA, xtable, data.table, matrixStats, RColorBrewer, fields, animation. By default, the missing packages will try to be automatically installed when compiling EPV_demo.Rnw

We would like to acknowledge STATS, LLC for consenting the inclusion of a full-game data sample.